As per the Turing test, a computer can be said to be intelligent if it can achieve human-level performance in all cognitive tasks, sufficient to fool an interrogator.
Defining AI into two categories
- Narrow AI – Particularly good at mastering one specific task like playing chess.
- Wide AI – Becoming more popular, a computer that can drive a car, play chess and detect fraudulent bank transactions. AI that can do more than one task better than a human or on a human-level.
One approach to AI is called Machine Learning, and is the most popular and widely used at the moment. It is a way of allowing machines to learn from their mistakes and improve on them without the input of a human handler. There are currently three main approaches to it. These are:
- Learning by labelled example: supervised learning. An example of this would be an email spam detector. It is usually amazingly effective if you have lots of examples or training data to feed into the AI.
- Discovering patterns: unsupervised learning. An example of this would be data clustering. It is difficult to implement in practice, but very useful if you lack labelled examples.
- Feedback right/wrong: reinforcement learning. An example would be learning to play chess by winning or losing. This method is very effective in some domains, and so it is becoming more important.
Another department of AI that is growing in popularity and is directly linked with how your online marketing is delivered to potential clients is Neural Networks. These are programs that are loosely based on how the biological/human brain works.
What Are Neural Networks Good for?
- Pixels: They can take pixels and tell you what it is. This is what is visible in Google’s Image search all the time.
- Audio: They can take audio in waveform, break it down and try to understand what this is. This is how a lot of the Google technology works at the present moment with Voice Search.
- Docs + Query: It can take a query plus a multitude of documents and tell you what the preferred version is. This is how Google’s text search works at the moment.
- Translation: Translate one language to another. This is the work that we are currently seeing with Google Translate and a whole lot of other language tutorial apps on the market.
Neural Networks, Machine Learning and Internet Search
Back in 2015, Google released an impressive machine learning algorithm known as RankBrain, while recently introducing machine learning and AI to its Android devices in the form of Tensorflow.
Google Rankbrain & Tensorflow
This algorithm functions alongside the Hummingbird update, helping Google understand the semantic meaning behind complex user queries. Released in the middle of 2015, RankBrain helps Google understand more complex user queries by reducing them down to more decipherable chunks. The AI component emerges because it can learn to make connections between words and phrases that aren’t immediately logically connected.
This is helping Google improve on its Voice Search service by trying to understand context and showing the best results when returning results to the user.
TensorFlow is an open-source software library custom made to make it easier for developers to produce machine learning programs. Google is putting AI in the hands of more developers and users.
Object Recognition in Images and Videos
This is also true for image recognition on Facebook, where Facebook recognises the user or the Facebook Filters, that are just a clone of Snapchat filters to be honest. With all the visual media being produced currently, advancements in this field can only get better with AI.
So now you know about the most important AI advancements related to businesses in my view, but what do they mean for your business? Well, there is no simple answer, but here are some important things to consider.
Smarter Search & Sentiment
With the vast amount of information being produced on internet, its going to get harder to stand out without a big budget. The best way is for your organisation to have AI in mind when optimising online campaigns and assets for search users. This means that it is going to be important to be whitelisted by search engines as well as being contextually relevant for the query you wish to rank for.
Whether it is a service or product, unless you are spending big money, all your information will have to be optimised. This can be achieved by “standing in the customer’s shoes” when approaching marketing.
Better Targeting for Ads
With technology companies slowly cracking the code of computers understanding the nuances of language, Small Businesses can expect better segmentation and targeting for their ads.
Instead of the old way of marketing, which involved focusing on who to show ads to and when to show the ads; now businesses can focus on what messages to show to the audience. This creates powerful ads specific to the target audience.
For example, Natural Language Processing (Cognitive Machine Learning technology) can allow for the matching of ads with individuals based on context, rather than keywords. With the large data available on users right now, its going to be easier to tailor your messaging to potential customers.
Chatbots and Messaging
Messaging platforms are everywhere we look nowadays. Whatsapp for friends, Facebook Messenger for or the in-app messenger on your website to talk to customers. If you don’t have one, you are already lagging behind. Chatbots allow you to prep AI to talk to consumers instead of having a human reply. This opens the door to 24/7 customer service – something that your consumers would surely appreciate.
Final Advice for Small Businesses looking into AI
- Firstly, make sure that your website is always adhering to search engine best practising by constantly optimising the SEO of your website. This will ensure that your website is white-listed and is eligible for RankBrain to match it to users who are more likely to convert.
- Take time to understand the audience you are making campaigns for, as you no longer have to rely on just keywords with your Paid Search campaigns. Understand the context in which the target audience might appreciate your ad more as they will be more likely to convert.
- Make sure to shoot good quality videos and images, so that they are better interpreted by these AI algorithms. Its no secret that visual media consumption is flourishing at the moment and this is the time to test out what works and what doesn’t – with all the tech available. Don’t be afraid to use filters and memes for campaigns.